The comparison everyone makes, and why it misleads
Most people put the price of a robot next to a yearly salary and conclude one of two things: that the robot is absurdly expensive, or that it pays back in a year. Both are wrong, for the same reason. The robot's price says nothing about how many hours it works, and the salary is a fraction of what a worker actually costs you.
The number that decides the case is simpler and harder: what does it cost to get one unit of work done, on this task, on this shift pattern, by a person and by a robot?
What an hour of labor really costs
Start with the wage for the job and the shift, including the night and weekend premiums. Then add what most people forget:
- Employer costs on top of the wage: taxes, insurance, benefits. Typically another twenty to forty percent, depending on the country.
- The agency's cut, if the shift is staffed through an agency. Commonly thirty to fifty percent on top.
- Overtime paid to cover the shifts nobody could be found for.
- Turnover: recruiting, training, and the weeks a new hire is slower than the person they replaced, spread over how long people actually stay.
- Absence, and the cost of covering it.
Add it up and divide by the hours people actually work, not the hours you pay for. Almost everyone who does this for the first time finds a number well above the wage they had in their head.
What an hour of robot really costs
- The robot itself: a monthly fee if you rent its time, or the purchase price and financing spread over its working life.
- Getting it working: fixtures, connecting it to your warehouse or labor system, safety measures. Spread over the deployment.
- Keeping it working: monitoring, maintenance, spare parts, the people watching it remotely.
- Electricity and charging. Small, but not nothing.
- Downtime: charging breaks and unplanned stops, which cut the hours it actually works.
Divide the yearly total by the hours it actually works. This is where the answer swings: nearly all of a robot's cost is fixed, so the same robot costs about three times as much per hour on one shift as it does on three.
Then compare on the work, not the hour
A robot does not match a person's pace on day one. So divide each side's cost per hour by how many units per hour it actually does on this task, using rates you have measured, not rates from a brochure. And allow for the ramp: the first weeks are slower.
A worked example, with round numbers
These figures are made up to show the mechanics. In a consultation, your own numbers go in their place.
| One shift | Three shifts | |
|---|---|---|
| Robot's yearly cost, all in | 100 | 100 |
| Hours it works per year | 1,800 | 5,400 |
| Robot cost per hour worked | 0.056 | 0.019 |
| Robot's pace as a share of a person's | 70% | 70% |
| Robot cost per unit of work, if a person's is 1.0 | about 1.6 | about 0.55 |
Read the last line. Same robot, same task: on one shift it costs sixty percent more than a person per unit of work; on three shifts it costs about half. That is why the shift pattern is one of the first things we ask, and why we build the case per task and per site, never per robot.
Where it usually works, and where it usually does not
With today's robots the case closes first where three things are true at once: the site runs more than one shift, the task is structured handling with a measured human rate, and the labor for it is expensive, scarce, or comes through an agency. Where one of the three is missing the case is usually not there yet. We will tell you so, because a pilot that cannot pay is not worth anyone's time.
